OVERCOMING CHILDREN’S GENDER-TYPED PREFERENCES FOR MUSICAL INSTRUMENTS: AN INTERVENTION STUDY A.C. Harrison and S.A. O’Neill, Department of Psychology, Keele University Introduction Social learning theory (e.g., Mischel, 1970) stresses the importance of observational learning in the acquisition of gender-typed behaviour. Accordingly, exposure to counter-stereotypic role-models has often been used to modify children’s gender-typed preferences (e.g., toy, activity, occupations), (Katz, 1986). Findings have, however, been mixed. Extensive exposure altered both young children’s stereotypes and verbally stated preferences (Koblinsky & Sugawara, 1984), but another sustained, 10 week intervention programme did not produce any significant differences in the gender-typed occupational preferences of kindergarten children (Weeks & Porter, 1983). Results have been similarly inconclusive for short-term interventions. DeLeo, Moely, and Sulzer (1979) reported provision of role-models to be effective in modifying the gender-typed beliefs of 4-6 year olds, whereas television presentation of counter-stereotypic role-models met with little success with either 7 year olds (Drabman, Robertson, Patterson, Jarvie, Hammer, & Cordua, 1981) or 1012 year olds (Williams, LaRose, & Frost, 1981). Huston (1983) argued that brief exposure to one model is unlikely to produce powerful effects because whilst this strategy might influence children’s stated beliefs about what is masculine or feminine, it does not, however, lead to changes in attitudes or affective reactions to genderinconsistent behaviour. 1 Studies have been most successful with 4-10 year olds. Weeks and Porter (1983) proposed that attempts to influence children’s gender-typed beliefs and preferences may be more successful when such stereotypes become more flexible. Thus, the present research focused on 7-8 year olds, the age at which gender-typed beliefs become more flexible (e.g., Carter & Patterson, 1982). This is also the age at which the majority of children in British schools are first offered the opportunity to learn a musical instrument (the domain focused upon in the research). A growing body of research suggests that children have well-established gender-stereotypes regarding musical instruments (e.g., Abeles & Porter, 1978; Delzell & Leppla, 1992; Harrison & O’Neill, in press; O’Neill & Boulton, 1996), and that these stereotypes are closely related to their preferences for musical instruments (Harrison & O’Neill, in press; O’Neill & Boulton, 1996). However, very little research has been conducted to develop intervention programmes designed to help children overcome gender-stereotyped beliefs and self-imposed restrictions on their choice of musical instruments. The need to address gender-typed beliefs in other domains such as sport, science, and mathematics has been recognised and attempts made to challenge such beliefs. It is important that issues regarding gendered participation in music are similarly addressed. For the most part, research has focused on establishing the existence, rather than challenging the maintenance of gender-stereotyped beliefs about musical instruments. Many researchers have, however, stressed the need to do so, particularly by the use of role-models. For example, Kelly (1997) suggested that presenting counter-stereotypic role-models may be “used to encourage the performance of gender-specific instruments by the non-traditional gender” (p. 54). 2 Extant research (e.g., Bruce & Kemp, 1993; Pickering & Repacholi, 1998; Tarnowski, 1993) suggests that the provision of counter-stereotypic role-models may reduce adherence to gender-stereotypes in expressing preferences for musical instruments. However, methodological limitations of these studies have meant that it is difficult to assess the efficacy of such an approach. The majority of studies have taken no measurement of initial preference, which means that researchers could not conclude that a preference change occurred, as many authors of these studies have suggested. No studies in this area have examined any enduring effects of intervention strategies. The present study aimed to further our understanding by conducting a short-term longitudinal study to investigate the effect of exposing 7-8 year olds to instruments played by either stereotypic or counter-stereotypic role-models. Research Questions The key questions were: 1. Do 7-8 year old girls’ and boys’ preferences for musical instruments change after they are presented with role-models which either agree or disagree with the children’s stereotypes? 2. To what extent do any changes in children’s instrument preferences remain 7 months later? Based on the findings of Bruce and Kemp (1993), we expected that girls would show greater preference for ‘masculine’ instruments (trumpet, guitar, drums) after observing female musicians playing those instruments, and that boys would indicate stronger preference for ‘feminine’ instruments (piano, flute, violin). Method 3 Participants Participants were 357 children (M = 185, F = 172) aged 7 and 8 years (Mean age = 8:1 years, SD = .29, range 7:7 to 9:3) attending junior schools in the south-west region of England. Twelve schools participated in the study, forming three clusters of four schools. Schools were selected to include rural and suburban catchment areas with pupils from working and middle class families. This age group was targeted because the majority of children had not yet participated in formal instrumental instruction, but would be offered the opportunity to begin learning an instrument at school within the next year. Measures A full description of measures is provided by Harrison and O’Neill (in press). Individual interviews were designed to obtain information about the children’s preferences for, and gender-stereotyped beliefs about, specific musical instruments. Participants were presented with line drawings (without performers) and audiocassette recordings of ‘masculine’ (trumpet, guitar, drums) and ‘feminine’ (piano, flute, violin) instruments. They were asked to rank-order the six instruments according to their preferences, from the one they would most like to learn to play to the one they would least like to learn to play. Children’s gender-stereotyped beliefs about the instruments were assessed by asking children to indicate whether the six instruments would be played by girls, boys, or both boys and girls. Children also completed a classroombased measure of instrument preferences. This comprised a single sheet with the same line-drawings of the instruments as used in the interviews. Participants were asked to circle and number the instruments in order of preference. 4 Intervention concerts The research design involved presenting a series of short demonstration concerts to two of the three school clusters during the research period. A full description of the intervention concerts is provided by Harrison and O’Neill (in press). The concerts took place during the same week, after completion of Time 1 interviews. Schools were randomly assigned into three groups. Cluster 1 schools received concerts played by gender-consistent role-models (e.g., female musician playing piano, male musician playing guitar). Cluster 2 received concerts played by gender-inconsistent role-models (e.g., female musician playing guitar, male musicians playing piano). The third group of schools did not receive concerts. All musicians were peripatetic music teachers and competent performers. The same music was performed to all schools and was selected to contrast in both tempo and musical style. Procedure Measures were administered on three occasions. At Time 1, instrument preferences and gender-stereotyped beliefs about those instruments were assessed during individual interviews. 234 participants were interviewed by the female researcher, and 123 children were interviewed by a male research assistant. Preferences were also assessed with a classroom-based measure. Intervention concerts were performed at two of three clusters of schools approximately one month after Time 1. Instrument preferences were measured using a classroom-based measure immediately after the concerts (Time 2) and approximately seven months after the intervention concerts (Time 3). All follow-up (Time 3) interviews were conducted by the female researcher. 5 Prior to the Time 1 interviews, the researcher was introduced to the children in the classroom by the class teacher. The children were informed about the study and reassured that no-one in the school would be told about their replies. The children's consent to participate was obtained by asking them to raise their hands if they were happy to help the researcher in "trying to find out what girls and boys think about some activities that children do". Only one child declined to participate. Parents were also contacted by letter prior to the study and invited to alert the relevant school if they did not wish for their child to participate in the study. Interviews took place at the selected schools in a quiet place, usually in the corridor or school library. Prior to each individual interview, in order to facilitate honest answers and satisfy ethical guidelines, participants were again told that their replies would remain confidential with respect to the adults and children in their school. They were told that the interview was not a test and that there were no right or wrong answers, and that they may stop at any time if they so wished. The classroom-based measure of instrument preferences was administered when all interviews had been completed (Time 1). Immediately following the concerts (Time 2), the children returned to their classrooms and their instrument preferences were assessed using the same classroom-based measure. For schools in Clusters 1 and 2, this was administered immediately following the concerts. Cluster 3 schools (control schools) were visited by the research assistant approximately one month after Time 1 and the same measure administered. The follow-up phase of the study (Time 3) was conducted approximately seven months after the intervention concerts. We had aimed to commence the followup phase six months after the concerts, but this was not possible due to constraints of the schools’ time-tables. Prior to the Time 3 interviews, the researcher was re- 6 introduced in the classroom by the class teacher. The children were reminded about their previous interviews. Interviews were conducted in a similar manner to the Time 1 interviews. Results Instrument preferences 87.4% of the children were present at Time 3 (n = 312, M = 155, F = 157). Harrison and O’Neill (in press) reported that girls liked piano and flute less when they had seen a man play the instruments. Boys liked guitar less when they had seen a woman play the instrument. Both girls and boys and liked drums more and violin less regardless of the gender of the musician they had observed playing these instruments. The present study aimed to investigate whether there was a lasting effect of providing intervention concerts on participants’ gender-typed instrument preferences. Tables 1 (females) and 2 (males) shows the mean rank given for each of the six instruments by participants from each of the three school clusters at Time 1, Time 2 and Time 3. Data are taken from the classroom-based measure. Six repeated-measures ANOVAs were carried out on each instrument separately to investigate any lasting effect of the intervention concerts on female, and separately, male participants’ preferences. How much male, and separately, female participants liked each instrument were the dependent measures, with one betweensubject factor (school cluster) and one within-subject factor (three levels of time period – Time 1, Time 2, Time 3). 7 Female participants There was a main effect of Time for piano (F (2, 292) = 5.62, p < .005) and drums (F (2, 292) = 16.09, p < ,0005). A main effect of Cluster was found for piano (F (2, 292) = 3.07, p < .04) and guitar (F (2, 292) = 3.34, p < .04). There was also a Time x Cluster interaction for piano (F (4, 292) = 4.40, p < .002), violin (F (4, 292) = 3.97, p < .005), and drums (F (4, 292) = 4.39, p < .002). Post hoc one-way ANOVAs were conducted to compare the female participants’ preferences for piano, violin, drums and guitar for each of the school clusters at each time period (Time 1, Time 2, Time 3) in order to investigate any lasting effect of providing gender-consistent or gender-inconsistent role models. Cluster 1 (gender-consistent role-models) For the violin, there was a main effect of Time, F (2, 80) = 4.85, p < .01. Female participants indicated less preference for violin at Time 3 (Mean = 3.56, SD = 1.52) than at Time 1 (Mean = 2.85, SD = 1.15), t (1, 40) = 2.80, p < .01). There was no significant difference in preference for the violin between Time 2 and Time 3. This indicates some lasting effect of the intervention concerts on the preferences of girls in Cluster 1 for the violin. There were no significant differences for any of the other instruments. Cluster 2 (gender-inconsistent role-models) There was a main effect of Time for piano (F (2, 100) = 10.66, p < .0005) and drums (F (2, 100) = 14.97, p < .0005). Female participants indicated more preference for piano at Time 3 (Mean = 2.59, SD = 1.78) than at Time 2 (Mean = 3.47, SD = 1.91), t (1, 50) = 3.24, p < .002. There was no difference between their liking for 8 piano at Time 1 and Time 3, indicating the decrease in the girls’ preference for piano immediately following the concerts did not last. Female participants liked drums more at Time 3 (Mean = 4.17, SD = 1.82) compared to Time 1 (Mean = 4.77, SD = 1.49), (t (1, 51) = 2.57, p < .02), but less at Time 3 (Mean = 4.14, SD = 1.82) than at Time 2 (Mean = 3.24, SD = 1.90), (t (50) = 2.71, p < .01), indicating that the increase in their liking for drums lasted to some extent seven months later. They liked violin less at Time 3 (Mean = 3.54, SD = 1.55) compared to Time 1 (Mean = 2.94, SD = 1.35), t (1, 51) = 2.34, p < .03), but there was no difference in how much they liked violin between Time 3 and Time 2. There was also a non-significant trend for flute to be ranked lower at Time 3 than at Time 1. Cluster 3 (no concerts) Surprisingly, participants ranked the violin more favourably at Time 3 than both Time 1(t (39) = 2.56, p < .02) and Time 2 (t (38) = 2.42, p < .02), even though they had received no intervention concerts. Male participants There was a main effect of Time for piano (F (2, 252) = 5.62, p < .005), and drums (F (2, 252) = 4.94, p < .01), and a Time x Cluster interaction for drums (F (4, 252) = 2.37, p = .052), and flute (F (4, 252) = 2.55, p < .04. Post hoc one-way ANOVAs were carried to compare preferences for flute, piano and drums for each of the school clusters at each time period (Time 1, Time 2, Time 3). We also compared preferences for violin and guitar as male participants’ preferences for these instruments altered following the concerts and we wanted to examine if this change persisted seven month later. 9 Cluster 1 (gender-consistent role-models) There was a main effect of Time for piano (F (2, 252) = 3.94, p < .03) and drums (F (2, 252) = 3.95, p < .03). Male participants liked piano more (Mean = 3.12, SD = 1.60) and drums less (Mean = 2.37, SD = 1.37) at Time 3 compared to Time 2 (piano, Mean = 3.83, SD = 1.47, t (1, 40) = 2.59, p < .02; drums, Mean = 1.83, SD = 1.02, t (2, 40) = 2.19, p < .04. There was no significant difference in preferences for drums between Time 1 and Time 3, indicating that the increase in preference immediately following the intervention concerts did not last. There were no significant differences in preferences for any of the other instruments. Cluster 2 (gender-inconsistent role-models) Participants liked flute less (Mean = 4.60, SD = 1.42) at Time 3 compared to Time 2 (Mean = 4.08, SD = 1.54), t (1, 46) = 2.66, p < .02). There was also a nonsignificant trend for participants to indicate greater liking for violin at Time 3 than at Time 2, but less liking for guitar at Time 3 than at Time 1. There were no significant differences in rankings of the six instruments by male participants in the Cluster 3 schools (controls). Comparison of participants’ gender-stereotyped beliefs at Time 1 and Time 3 Harrison and O’Neill (in press) reported that male and female participants had similar ideas about which sex would play specific instruments at Time 1 – they considered piano, violin and flute were ‘for girls’ and trumpet, guitar and drums to be ‘for boys’. In the follow-up phase, we were interested in whether the children’s gender-stereotyped associations of instruments had altered since Time1 (seven 10 months after intervention concerts). We compared ‘flexibility’ (the number of initial ‘both’ responses) of gender-stereotypes at Time 1 and Time 3. We also examined any changes in participants’ gender-typed beliefs about masculine, and separately, feminine instruments, comparing the children’s forced-choice selections since the forced-choice format measures primarily knowledge of stereotypes rather than attitudes to gender-roles (Signorella, Bigler & Liben, 1993). Proportionally more female, and separately, male participants changed their stereotypes of ‘masculine’ (trumpet, guitar, drums) instruments from “for girls” to “for boys” (females, 27.4%; males, 28.9%) than changed their opinion from “for boys” to “for girls” (females, 15.0%), X2(1) = 5.87, p < .02; males, 10.9%, X2(1) = 18.99, p < .001. Thus, whilst stereotypes of instruments did not change greatly after the concerts, when children altered their stereotyped-beliefs about instruments, they tended to change their opinion to the ‘correct’, culturally accepted stereotype. Flexibility of gender-stereotypes Following a similar procedure to that used by Serbin, Powlishta, and Gulko (1993), a “flexibility” score was computed from the number of “both” responses given. “Flexibility” refers to the extent to which the child believes that a culturally gender-stereotyped activity or role is appropriate for both females and males (Serbin et al, 1993). Scores could range from 0 to 6. There was no significant difference between male and female participants’ flexibility scores at Time 1 or Time 3. However, male participants showed more flexibility at Time 3 (Mean = 2.88, SD = 1.33) than at Time 1 (Mean = 2.62, SD = 1.45), t (155) = 2.00, p < .0001. Similarly, female participants scored higher on flexibility at Time 3 (Mean = 3.08, SD = 1.34) than at Time 1 (Mean = 2.78, SD = 1.44), t (156) = 2.31, p < .0001. These findings 11 are consistent with those of Serbin et al (1993), who noted an increase in flexibility scores in second and third graders, the age group in the present study. We also compared female and male participants’ flexibility scores at the two time points within each school cluster. Male participants attending Cluster 1 (gender-consistent) schools scored higher on flexibility at Time 3 (Mean = 2.70, SD = 1.17) (t (42) = 2.02, p < .05) than at Time 1 (Mean = 2.26, SD = 1.26). There was a non-significant trend for female participants in Cluster 3 (control) schools to score higher on flexibility at Time 3 (Mean = 3.53, SD = 1.28) than at Time 1 (Mean = 3.13, SD = 1.53). This suggests that the overall increase in flexibility was attributable more to age-related change rather than any possible influence of intervention concerts. Discussion We reported previously that there was an immediate effect of providing counter-stereotypical role-models on boys’ and girls’ preferences for genderconsistent and gender-inconsistent instruments (see Harrison & O’Neill, in press, for discussion of results). The present study indicated some lasting influence of the intervention concerts on the children’s instrument preferences. Boys who saw a woman play guitar continued to like it less, and liked violin less regardless of whether they had seen a man or woman play it. The increase in their liking for drums did not last. Girls still liked violin and drums more (after seeing a man or woman play the instruments), and flute less (after seeing a man play it) seven months after the concerts. However, the decrease in girls’ liking for piano (after seeing a man play it) did not last. These results suggest that the decrease in liking for same-sex instruments did to some extent endure over seven months. Thus, we found more enduring effects of 12 providing role-models compared to other studies designed to challenge children’s gender-typed beliefs. However, our study suggests that exposing children to counterstereotypic role-models is not necessarily effective in either the short- or long-term. For example, girls actually liked flute less after seeing a man play it, and this effect was still evident seven months later. Other studies in the literature have reported negative effects of such strategies. For example, eighth grade boys expressed more stereotypical attitudes about women’s roles after watching counter-stereotypic advertisements compared to stereotypic commercials. Guttenbrag and Bray (1976) presented participants with films and stories about women with counter-stereotypic roles. Whilst female participants’ stereotypic attitudes decreased significantly, male participants in fifth and ninth grade showed more stereotypical attitudes following presentation of the role-models. Our data suggests that children were more likely to change their preferences for instruments rather than the stereotype itself. Thus, findings indicate that presenting counter-stereotypic role-models may be an ineffective strategy for challenging children’s gender-stereotyped beliefs. More research is needed before we will have a better understanding of the best way to present instruments to children if we are to assist them in overcoming the tendency to restrict instrument choice along grounds of gender. It may be that role-models should be presented separately in single-sex teams to overcome the negative effect we observed in our study (see also Matteson, 1991). In the mainstream gender literature, longer and more interactive intervention studies have been more successful (e.g., Koblinsky & Sugawara, 1978; Weeks & Porter, 1983). Despite various methodological problems, Tarnoswki’s (1993) study suggested that gender-neutral instrument presentation workshops (over a period of eight weeks) may be beneficial in encouraging children to hold gender-neutral beliefs about the 13 gendered nature of musical instruments. We propose that it may not be enough to merely expose children passively to musicians that either do or do not conform to their stereotypes, rather it may be necessary for children to engage actively in considering the role of gender in playing instruments through special programmes designed to raise their awareness and develop more gender-neutral attitudes. References Abeles, H. F., & Porter, S. Y. (1978). The sex-stereotyping of musical instruments. Journal of Research in Music Education, 26, 65-75. Bruce, R., & Kemp, A. (1993). Sex-stereotyping in children’s preferences for musical instruments. British Journal of Music Education, 10, 213-217. Carter, B., & Patterson, C. J. (1982). Sex roles as social conventions: the development of children‘s conceptions of sex-role stereotypes. Developmental Psychology, 18, 812-824. Delzell, J. K., & Leppla, D. A. (1992). Gender associations of musical instruments and preferences of fourth-grade students for selected instruments. Journal of Research in Music Education, 40, 93-103. DiLeo, J. C., Moely, B. E., & Sulzer, J. L. (1979). Frequency and modifiability of children‘s preferences for sex-typed toys, games, and occupations. Child Study Journal, 9(2), 141-159. Drabman, R. S., Robertson, S. J., Patterson, J. N., Jarvie, G. J., Hammer, D., & Cordua, G. (1981). Children‘s perceptions of media-portrayed sex roles. Sex Roles, 7(4), 379-389. Guttenbrag, M., & Bray, H. (1976). Undoing sex stereotypes: Research and resources for educators. New York: McGraw-Hill. 14 Harrison, A. C., & O'Neill, S. A. (in press). Children' gender-typed preferences for musical instruments: an intervention study. Psychology of Music. Huston, A. C. (1983). Sex-typing. In E. M. Hetheringon (Ed.), Handbook of child psychology: Socialization, personality and social development (Vol. 4, pp. 288-467). New York: Wiley. Katz, P. A., & Boswell, S. (1986). Flexibility and traditionality in children's gender roles. Genetic, Social, and General Psychology Monographs, 112(1), 103-147. Kelly, S. N. (1997). An investigation of the influence of timbre on gender and instrument association. Contributions to music education, 24(1), 43-56. Koblinsky, S. A., & Sugawara, A. I. (1984). Non-sexist curricula, sex of teacher and children’s sex-role learning. Sex Roles, 10(5/6), 357-367. Matteson, D. R. (1991). Attempting to change sex role attitudes in adolescents: explorations of reverse effects. Adolescence, 26(104), 885-898. Mischel, W. (1970). Sex typing and socialization. In P. H. Mussen (Ed.), Carmichael's handbook of child psychology (Vol. 2, pp. 3-72). New York: Wiley. O'Neill, S. A., & Boulton, M. J. (1996). Boys’ and girls’ preferences for musical instruments: a function of gender? Psychology of Music, 24, 171-183. Pickering, S., & Repacholi, B. (1999). The effect of musician gender on children‘s instrument preferences. Paper presented at the Society for Research in Child Development Biennial Meeting, Albuquerque. Serbin, L.A., Powlishta, K.K., & Gulko, J. (1993) The development of sex-typing in middle childhood. Monographs of the Society for Research in Child Development, 58. 15 Signorella, M. L., Bigler, R. S., & Liben, L. S. (1993). Development differences in children’s gender schemata about others: a meta-analytic review. Developmental Review, 13, 147-183. Tarnowski, S. M. (1993). Gender bias and musical instrument preference. Update: The Applications of Research in Music Education, 12(1), 14-21. Weeks, O. N. M., & Pryor Porter, E. (1983). A second look at the impact of nontraditional vocational role models and curriculum on the vocational role preferences of kindergarten children. Journal of Vocational Behaviour, 23, 6471. Williams, F., LaRose, R., & Frost, F. (1981). Children, television and sex-role stereotyping. New York: Praeger Publishers. 16